Politico: In the US, congressmen are massively using artificial intelligence to generate bills — the documents are “rife with errors”

The US Congressional lawyers are struggling to cope with legislative proposals created using artificial intelligence tools — many of them contain errors and are written very carelessly.

This was reported by eight current and former officials who work or have worked with the Office of Legislative Counsel of the US House of Representatives, writes Politico.

According to them, in many cases, projects generated by artificial intelligence are full of errors — due to the long work of lawyers on them, this can delay the adoption of laws and create the risk of lawsuits related to incorrectly cited legal norms or incorrect legal definitions.

The source noted that given the number of errors generated by AI tools, lawyers “are spending more time correcting AI-generated bills than they would if they were written from scratch”. Another problem with the bills is that they make fictional or incorrect references to existing laws and sections of the US Code.

Politicoʼs interlocutors spoke about another problem. The use of AI changes not only the quality of the texts, but also how deeply their authors understand the issues they are trying to resolve. In particular, the people who bring the document often cannot clearly explain what result they want to achieve.

However, there are currently no confirmed cases where the House or Senate voted on a bill created entirely by artificial intelligence.

For more news and in-depth stories from Ukraine, please follow us on X.